Great sensitivity and good control is needed to create a "programmable" or maybe a "controllable" power supply would be a better word. The measurement should be made across two terminals ( the ends of a wire). The current through the wire needs to be measured as well as the the voltage across the two ends. These two measurements should be done nearly as simultaneously as possible. I propose a pulse width modulated "power" control system. A verbal discription follows:

First, an estimate of the power output requirements. The MFMP project ha;s been using somewhere between 26 and 100 watts. So let's say the design power maximum should be 200 watts. If we say the maximum voltage (fixed?) in the supply is 24 volts, then we will need a primary supply of 24 volts and 10 amps. I bet we can buy this device already made and ready to use. Just plug it into the wall power socket.
The pulse width modulated controller can be based on www.digikey.com/product-detail/en/EA-XPR...A-XPR-003-ND/2410099 or similar. The onboard CPU resources contain several 16 bit pwm controllers. We would want to operate at about 500 hertz. or a little faster if possible. A 16 bit timer can be set up at a clock frequency of about ( clk=1/500/2^16 ) seconds. This gives a duty-cycle resolution of 2^16 or about 1 part in 32000. The pwm should drive a logic level MOSFET able to switch at least 20 amps with a source-drain voltage of greated than 50 volts. say a BUK9510-55A,127 from NXP. This is avalable from DIGIKEY for under $2.00.
We will also need a resistot as a current "sense" device. THis should be put below th mosfet switch in the circuit with the wire we are trying to monitor. A power resistor of 0.25 ohms should work well. A differential amplifiier will exract a voltage proportional to the circuit's current, but only a while the MOSFET is switched on.
At nearly the same time, another differential amplifier will sample the voltage across the interesting wire. At about 500,000 samples/second, this voltage should be captured about 2 or 3 microseconds later- nearly simutaneously.
The CPU will then calculate the power ( volts*amps*(duty cycle)). If the power is too low, then increase the duty cycle. If the power is too high, then decrease the duty cycle.
Both wires will need the same control circuit.
Some purists may argue this is not DC or exactly what Celani did or whatever. I present the strategy as merely a very accurate stategy for operation of a programmable "power" supply. With some careful construction we should be able to measure 200 watts to about 1 part in 32000 accuracy. This is good enought for most applications. The accuracy should be realized from a fraction of a watt to 200 watts.
The thermal time constant of most objects is much bigger than 1/500 of a second. If the object's time constant is greater than about the reciprocal of 1/5 or 1/6 of the sampling frequency, the object will not see any signifigant thermal pulses.

jdk

Hi JDK,

For basic ultra stable programmable and controllable PSU we are mined to use best in class off the shelf units like the Thurlby Thandar ones that donations allowed us to buy for the EU cell - they were very effective.

However, we want to launch 2 mini project in the Collaborate section

1. A simple wide band (at least taking into consideration Piantelli's patent insight) pulse/waveform generator with variable pulse width and amplitude.

2. A VERY HIGH POWER MOSFET (say the ones from STM incidentally!) based power amp/switch This would allow very high amps / volts (power) to be momentarily switch to the active wires for say a millisecond every second

This would allow us to test electrical shock and wave based triggering.

We would then want another mini project

3. Radio frequency generator and suitable emitter for the cell.

I just looked at some of the specs for the "Thurlby Thandar" power supplies. I don't know which one we're useing, but the update time seemed slow (40 millisec.) for both voltage and current. (80 msec?). I think we should build a PWM controller based on big MOSFETs and high voltage power supplies. If the resistance of the wires in the experiment are 15 ohms and we want to generate 100 amps current flow, then we will need a 1500 volt power supply. The duty cycle follows: let the target average (DC) watts = 50, then 50= 1500^2/15/duty estimates the duty cycle at 0.000333333. This would generate 100 amp pulses into the resistive load and keep the average power at 50 watts. This is about a 3000 hertz update rate for an average power of 50 watts with an instaneous current flow of 100 amps. Instaneous power would be 100^2*15. I^2*R . 15*10^4 watts peak 150,000 watts peak. If more peak power is needed, the voltage must be raised. I think there are no mosfets that will work in this voltage range. Maybe an IGBT or SCR or whatever (more research needed). A big capacitor can be used to accumulate the charge and then dumped through the resitive load.

I haven't had time to work out all the details yet, but I think the path is clear. Custom circuitry with very high voltage properties and fast switching capabilities.


jdk